curl currently only works with libpsl's built-in PSL database and thus it will slowly go out of date over time unless libpsl is updated with a newer built-in database. Some operating systems (linux distros) now rather provide an "publicsuffix-list" package that updates the PSL data, so if curl would just use the correct libpsl API it could be using more up-to-date PSL information.
This libpsl API: psl_latest, is available since libpsl 0.16.0 (Dec 2016)
psl_latest() allocates memory, loads the appropriate file contents into that dynamic memory and returns a pointer to a new "context" that needs psl_free() to get freed after use.
We thus need to cache the loaded results somewhere for some amount of time after use to make this bearable and not cause too much of a performance impact. We can't make it globally cached without also making it not thread safe.